Research On Social Attribute Aware Task Scheduling Strategy In Edge Computing

The rapid development of the internet of everything application service has spawned mobile edge computing,which is a key software and hardware support platform for the era of big data processing in the context of internet of everything.In the era of the internet of everything,the amount of data on edge devices is large,and the real-time data processing requirements are high.Under the existing cloud computing architecture,data processing brings new challenges to transmission bandwidth and computing load.The proposed edge computing architecture makes up for the shortcomings of cloud computing.The new data processing mode in edge computing can not only ensure short response time and high reliability,but also use special sensing technology to achieve more and more applications services migrate from cloud computing centers to network edge devices.If most of the data can be processed on the edge without uploading to the cloud computing center,this will greatly save transmission bandwidth,power consumption for data transmission,slow down the load on the cloud center,and improve data privacy protection.Firstly,the characteristics of mobile communication nowadays are introduced,and the basic concepts of mobile edge computing are introduced through traditional cloud computing technology.Furthermore,its representative characteristics,application scenarios,hotspot issues are introduced,and the current existing strategy of task scheduling is briefly analyzed.Firstly,a social attributes and server available resources-aware task scheduling strategies for edge computing is proposed.D2 D links are allowed to reuse and share the communication resources of adjacent cellular users in response to assisting task scheduling.Specifically,by reinforcement Q learning algorithm,the best server corresponding to each task is selected.By measuring the terminal activity and intimacy between terminal and the edge server,the social attribute is quantified as a standard of terminal selection.By power control,the effective transmission of the task is guaranteed while avoiding interference to cellular users.Numerical results verify that the proposed strategy can effectively schedule the tasks of sensor nodes and achieve the maximum load balancing of the network.Further more,a software-defined structure for edge computing tasks collaborative scheduling is proposed.Firstly,based on idle users and overloaded users,a two-way selection clustering algorithm is proposed,which combines the historical information of idle users and the similarity of interests with overloaded users to form a stable collaborative cluster.Then,a sub-task partitioning algorithm based on optimal delay is proposed to achieve the optimal overall delay while ensuring that the sub-tasks are completed simultaneously.Numerical results show that the proposed strategy can not only significantly save the data transmission bandwidth,but also achieve the optimal time delay while ensuring the stability of the cooperative cluster.